CN112766580B - Multi-source precipitation product fusion method based on dynamic heuristic algorithm - Google Patents

Multi-source precipitation product fusion method based on dynamic heuristic algorithm Download PDF

Info

Publication number
CN112766580B
CN112766580B CN202110096858.2A CN202110096858A CN112766580B CN 112766580 B CN112766580 B CN 112766580B CN 202110096858 A CN202110096858 A CN 202110096858A CN 112766580 B CN112766580 B CN 112766580B
Authority
CN
China
Prior art keywords
precipitation
magnitude
data
simulation
data source
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202110096858.2A
Other languages
Chinese (zh)
Other versions
CN112766580A (en
Inventor
尹家波
郭生练
于兵
何绍坤
田晶
邓乐乐
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan University WHU
Original Assignee
Wuhan University WHU
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wuhan University WHU filed Critical Wuhan University WHU
Priority to CN202110096858.2A priority Critical patent/CN112766580B/en
Publication of CN112766580A publication Critical patent/CN112766580A/en
Application granted granted Critical
Publication of CN112766580B publication Critical patent/CN112766580B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N7/00—Computing arrangements based on specific mathematical models
    • G06N7/01—Probabilistic graphical models, e.g. probabilistic networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Human Resources & Organizations (AREA)
  • Data Mining & Analysis (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Evolutionary Computation (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Operations Research (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Marketing (AREA)
  • Quality & Reliability (AREA)
  • Game Theory and Decision Science (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Development Economics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Mathematical Optimization (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computational Linguistics (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Health & Medical Sciences (AREA)
  • Educational Administration (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Analysis (AREA)
  • Computational Mathematics (AREA)

Abstract

The invention provides a multi-source precipitation product fusion method based on a dynamic heuristic algorithm, which comprises the following steps: collecting limited site observation data, satellite inversion and reanalysis precipitation data set of a scarce data area; correcting each data set by adopting a local magnitude scaling method and an equal rate correction method, and obtaining a correction coefficient of each station; describing the capturing capacity of each data source to precipitation events through the assigning function, and calculating the state weight with space-time dynamic attributes; considering the simulation capability of each data source to the precipitation magnitude, and calculating dynamic magnitude weight based on the cuckoo algorithm; and (3) dynamically mapping correction coefficients, state weights and magnitude weights of all stations to the same spatial resolution by adopting a common Krigin interpolation method, and obtaining a long series of grid quantitative precipitation products through correction and data fusion. The advantages of the rainfall data sources are effectively fused, the system deviation of the rainfall data sources is corrected, and a reference basis is provided for watershed hydrological simulation and water resource planning.

Description

Multi-source precipitation product fusion method based on dynamic heuristic algorithm
Technical Field
The invention belongs to the technical field of multi-source data fusion, and particularly relates to a fusion method of a multi-source precipitation product based on a dynamic heuristic algorithm.
Background
High-quality quantitative precipitation data are important basic data for disaster prevention and control early warning, agricultural production management, ecological protection, watershed hydrological simulation and hydraulic engineering planning and design. Traditional precipitation data mainly depend on site observation, but a meteorological station network is usually low in density and uneven in space arrangement, so that the space-time change characteristic of precipitation is difficult to accurately reflect, and the engineering application requirements such as high-precision hydrological simulation cannot be met.
In recent years, satellite telemetry and data inversion algorithms are rapidly developed, precipitation quantitative observation products based on satellite remote sensing inversion have wider coverage range and higher space-time resolution, the defect of insufficient arrangement of meteorological stations is effectively overcome, and new data reference is provided for data-free areas. Meanwhile, as human observation means and data assimilation technologies become mature day by day, students perform quality control on observation data from various sources (ground, ships, radiosonde, anemometry balloons, airplanes, satellites and the like), and propose a data assimilation technology for numerical weather forecast to reconstruct a long-term historical climate process, namely a so-called reanalysis data set, which assimilates the numerical weather forecast and a large amount of ground observation data and satellite remote sensing information, and has the advantages of high spatial and temporal resolution precision, long time span and the like.
However, the method is limited by the influences of remote sensing precision, inversion algorithm, numerical prediction mode, assimilation scheme and the like, and both the satellite and the reanalysis precipitation product have large system deviation. The domestic scholars propose a multi-source data fusion method which can effectively improve the estimation precision of precipitation data. However, the following problems are common to the existing studies: (1) the space-time heterogeneity of precipitation is generally ignored, and the dynamic characteristics of fusion parameters are less considered; (2) systematic deviations of the precipitation data sources cannot be corrected, in particular the frequency and magnitude of the precipitation are not corrected simultaneously. Generally, the existing research cannot effectively integrate the advantages of various precipitation data sources, and limits the basin planning, management and guidance of the operation scheduling of hydraulic engineering.
Disclosure of Invention
The invention aims to provide a fusion method of multi-source precipitation products based on a dynamic heuristic algorithm aiming at the defects of the prior art, the method effectively fuses the advantages of various precipitation data sources, and provides an important reference basis with strong operability for basin hydrological simulation and water resource planning.
In order to solve the technical problems, the invention adopts the following technical scheme:
a multi-source precipitation product fusion method based on a dynamic heuristic algorithm comprises the following steps:
step 1, collecting limited site observation data of a scarce data area, performing satellite inversion and re-analyzing a precipitation data set;
step 2, correcting the data collected in the step 1 by adopting a local magnitude scaling method and an equal rate correction method, and obtaining correction coefficients of all stations;
step 3, describing the capturing capacity of each data source to precipitation events through the assigning function, and calculating the state weight with space-time dynamic attributes;
step 4, considering the simulation capability of each data source on the precipitation magnitude, and calculating the dynamic magnitude weight based on the cuckoo algorithm;
and 5, adopting a common Krigin interpolation method to dynamically map the correction coefficients, the state weights and the magnitude weights of all the stations to the same spatial resolution, and obtaining a long series of grid quantitative precipitation products through correction and data fusion.
Further, step 2 further comprises the following sub-steps:
step 2.1, correcting the precipitation occurrence frequency of each data source month by adopting a local magnitude scaling method according to the data of each ground station;
and 2.2, calculating correction factors of different months of each station based on an equal rate correction method, and obtaining the precipitation magnitude corrected by each data source.
Further, the specific operation method of step 2.1 is as follows: firstly, calculating precipitation occurrence frequency of each station based on measured data, and substituting the frequency into a frequency distribution function of each simulated precipitation magnitude, so as to obtain a simulated threshold of each data source; finally, by analyzing the difference between the simulation threshold and the judgment threshold and adopting a magnitude scaling method, the precipitation occurrence probabilities of the simulation series and the actual measurement series in different months are consistent, and the specific method comprises the following steps: when the simulation threshold is larger than the judgment threshold, correcting precipitation magnitude exceeding the judgment threshold but smaller than the simulation threshold as the judgment threshold in magnitude scaling; when the simulation threshold is smaller than the judgment threshold, the precipitation magnitude lower than the judgment threshold but higher than the simulation threshold is corrected as the judgment threshold in magnitude scaling.
Further, the specific operation method of step 2.1 is as follows:
for the jth ground station, note
Figure BDA0002914627810000021
And calculating the correction factor of the mth month for the simulation data source of the mth set by adopting the following formula:
Figure BDA0002914627810000022
in the formula: d represents the total observed days of the mth month of the station j, and D represents the time sequence of the daily precipitation series;
Figure BDA0002914627810000023
the measured precipitation magnitude of the site j on day d of the mth month;
Figure BDA0002914627810000024
the precipitation magnitude corrected by a local magnitude scaling method is adopted for the nth set of simulation data sources of the station j on the d day of the m month;
by adopting the correction factor, the corrected magnitude of the s-th set of analog data source can be obtained as follows:
Figure BDA0002914627810000031
in the formula:
Figure BDA0002914627810000032
and (4) representing the precipitation magnitude corrected by the local magnitude scaling method on the d th day of the m month by the s set of simulation data source of the station j, and further correcting the precipitation magnitude by using an equal rate correction method.
Further, the specific method of step 3 is:
the space-time heterogeneity of the precipitation events is fully considered, and a scoring function is adopted
Figure BDA0002914627810000033
And (3) representing the detection capability of each corrected analog data source on the precipitation event:
Figure BDA0002914627810000034
in the formula:
Figure BDA0002914627810000035
the rainfall magnitude corrected by using the LOCI method on the d th day of the m month by the s set of simulation data sources representing the station j is further corrected by using an equal rate correction method;
Figure BDA0002914627810000036
the measured precipitation magnitude of the ith set of simulation data source of the characterization station j in the mth month and the d day; roCharacterizing a precipitation threshold; l (u) is an indicator function, and when u is more than or equal to 0, L (u) is 1; whereas l (u) is 0;
then, the state weight omega of the mth set of simulation data source on the d th day of the mth month is calculated by adopting the following formula1,s(d)：
Figure BDA0002914627810000037
In the formula:
Figure BDA0002914627810000038
for the indication function, K is the number of the fused precipitation data sources; when in use
Figure BDA0002914627810000039
Figure BDA00029146278100000310
On the contrary, the method can be used for carrying out the following steps,
Figure BDA00029146278100000311
further, step 4 further comprises the following sub-steps:
4.1, optimizing the memory length based on a dynamic Bayesian mode averaging method;
and 4.2, calculating the dynamic magnitude weight of each station based on the cuckoo algorithm.
Further, the specific method in step 4.2 is as follows:
for each ground station, the minimum sum of squared deviations is used as a target function, and the cuckoo algorithm is adopted to calculate the dynamic magnitude weight of each data source:
Figure BDA0002914627810000041
in the formula: w is as(d) Simulating the magnitude weight of the data source dynamics at day d of the mth month for the mth set,
Figure BDA0002914627810000042
the s set of simulation data sources representing the station j represent the precipitation magnitude corrected by the equal rate correction method in the step 2 on the day of the mth month;
Figure BDA0002914627810000043
the measured precipitation magnitude of the s-th set of simulation data sources representing the station j on day of the mth month; and N is recorded as the memory length, and K is the fused precipitation data quantity.
Further, the specific method in step 5 is as follows:
step 5.1, adopting a common Kriging interpolation method to interpolate the correction coefficients of all the stations in the step 2, the state weight in the step 3 and the magnitude weight in the step 4 into a grid scale;
step 5.2, for each grid, based on the interpolated precipitation threshold value and the correction coefficient, adopting a local magnitude scaling method and an equal rate correction method to obtain a plurality of corrected sets of simulated precipitation data sources;
step 5.3, for each i grids, based on the state weight obtained by interpolation, adopting
Figure BDA0002914627810000044
Calculating the precipitation magnitude I of the ith set of simulated precipitation data source of the ith grid on the d dayi,dWherein
Figure BDA0002914627810000045
Simulating the precipitation magnitude of the precipitation data source on the d day for the s set of the ith grid corrected in the step 5.2, wherein K is the number of the fused precipitation data;
step 5.4, for each i grids, based on magnitude weight obtained by interpolation, adopting
Figure BDA0002914627810000046
Calculating the precipitation magnitude of the ith set of simulated precipitation data source of the ith grid on the d day
Figure BDA0002914627810000047
Step 5.5, adopting the precipitation magnitude I of step 5.3i,dJudging whether precipitation event occurs, if Ii,dIf the rainfall is less than 1, the precipitation event does not occur, and the fused precipitation value is determined as Ii,d(ii) a Otherwise, if the precipitation event occurs, the fusion precipitation value is calculated in the step 5.4
Figure BDA0002914627810000048
Compared with the prior art, the invention has the beneficial effects that: the advantages of satellite remote sensing, atmospheric reanalysis and ground meteorological station data are fully exerted, the defects that the observation space resolution is low and the ground observation station series are short in the satellite inversion and reanalysis technology are overcome, long-series rainfall data are obtained through the data fusion and deviation correction technology, and the fused rasterized long-series rainfall data provide important reference basis with strong operability for basin hydrological simulation and water resource planning; the invention considers the space-time heterogeneity of the precipitation, considers the dynamic characteristics of the fusion parameters, corrects the system deviation of each precipitation data source, particularly corrects the occurrence frequency and magnitude of the precipitation, and is beneficial to basin planning, management and guidance of the operation scheduling of the hydraulic engineering.
Drawings
Fig. 1 is a detailed flow chart of a fusion method of a multi-source precipitation product according to an embodiment of the invention;
FIG. 2 is a schematic diagram of a posterior probability density function of a satellite precipitation data source according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the following embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that the embodiments and features of the embodiments may be combined with each other without conflict.
The present invention is further illustrated by the following examples, which are not to be construed as limiting the invention.
The invention provides a multi-source precipitation product fusion method based on a dynamic heuristic algorithm, which comprises the steps of firstly collecting limited station observation data of a scarce data area, performing satellite inversion and re-analyzing a precipitation data set; correcting each data set by adopting a local magnitude scaling method and an equal rate correction method, and obtaining a correction coefficient of each station; then, describing the capturing capacity of each data source to precipitation events through an assigning function, and calculating the state weight with space-time dynamic attributes; then, considering the simulation capability of each data source on the precipitation magnitude, and calculating the dynamic magnitude weight based on the cuckoo algorithm; and finally, dynamically mapping correction coefficients, state weights and magnitude weights of all stations to the same spatial resolution by adopting a common Krigin interpolation method, and obtaining a long series of grid quantitative precipitation products through correction and data fusion, wherein the detailed flow is shown in figure 1.
The technical scheme of the invention is further explained in detail by the following embodiments and the accompanying drawings:
step 1, collecting limited site observation data of a scarce data area, performing satellite inversion and re-analyzing a precipitation data set;
in the embodiment, 4 satellite inversion precipitation products are adopted, namely IMERG Final products of GPM satellites, TMPA 3B42V7 products of TRMM satellites, PERSIANN-CDR products developed by European university of California and MSWEP-V2 data sets developed by Princeton university of California.
The reanalysis precipitation data set adopted in the embodiment is a fifth generation reanalysis weather product ERA5 of the European middle-term weather forecast center; the horizontal resolution of the hourly analysis field of the data set is 31km, 137 layers are vertically layered, and the top layer reaches the height of 0.01 hPa; the ERA5 adopts a Cycle31r2 model version of a comprehensive forecasting system, based on spectral harmonic resolution T255, and interpolates simplified Gaussian grid (N128) data to grids with different resolutions of 0.25-2.5 degrees and the like by a bilinear interpolation technology, so that the data is one of global reanalysis data with the highest space-time resolution at present.
And 2, correcting each data set by adopting a local magnitude scaling method and an equal rate correction method, and obtaining a correction coefficient of each station. This step further comprises the sub-steps of:
step 2.1, correcting the precipitation occurrence frequency of each data source month by adopting a Local intensity scaling (LOCI) method according to the data of each ground station;
in order to carry out deviation correction on the satellite remote measurement and reanalysis precipitation data source, firstly extracting grids corresponding to all stations so as to obtain corresponding simulated precipitation series; in this embodiment, 1mm is set as a judgment threshold of a precipitation event, the precipitation occurrence frequency of each station is calculated based on actual measurement data, and the frequency is substituted into a frequency distribution function of each simulated precipitation magnitude, so as to obtain a simulated threshold of each data source; finally, by analyzing the difference between the simulation threshold and the judgment threshold and adopting a magnitude scaling method, the precipitation occurrence probabilities of the simulation series and the actual measurement series in different months are consistent, and the specific method comprises the following steps: when the simulation threshold is larger than the judgment threshold, correcting precipitation magnitude exceeding the judgment threshold but smaller than the simulation threshold as the judgment threshold in magnitude scaling; when the simulation threshold is smaller than the judgment threshold, the precipitation magnitude lower than the judgment threshold but higher than the simulation threshold is corrected as the judgment threshold in magnitude scaling.
Step 2.2, based on an equal rate correction method, calculating correction factors of different months of each station, and obtaining the precipitation magnitude after correction of each data source;
for the jth ground station, note
Figure BDA0002914627810000061
And calculating the correction factor of the mth month for the simulation data source of the mth set by adopting the following formula:
Figure BDA0002914627810000062
in the formula: d represents the total observed days of the mth month of the station j, and D represents the time sequence of the daily precipitation series;
Figure BDA0002914627810000063
the measured precipitation magnitude of the site j on day d of the mth month;
Figure BDA0002914627810000064
simulating the precipitation magnitude of a data source of the station j on the d th day of the m month after the data source is corrected by using the LOCI method;
by adopting the correction factor, the corrected magnitude of the s-th set of analog data source can be obtained as follows:
Figure BDA0002914627810000065
in the formula:
Figure BDA0002914627810000066
and (4) representing the precipitation magnitude corrected by the LOCI method on the d th day of the m month by using the simulation data source of the station j.
And 3, describing the capturing capacity of each data source to the precipitation event through the assigning function, and calculating the state weight with the space-time dynamic attribute.
The space-time heterogeneity of the precipitation events is fully considered, and a scoring function is adopted
Figure BDA0002914627810000071
And (3) representing the detection capability of each analog data source corrected in the step (2) on the precipitation event:
Figure BDA0002914627810000072
in the formula:
Figure BDA0002914627810000073
the rainfall magnitude corrected by using the LOCI method on the d th day of the m month by the s set of simulation data sources representing the station j is further corrected by using an equal rate correction method;
Figure BDA0002914627810000074
the measured precipitation magnitude of the ith set of simulation data source of the characterization station j in the mth month and the d day; roThe precipitation threshold is characterized, and the value of 1mm is taken in the embodiment; l (u) is an indicator function, and when u is more than or equal to 0, L (u) is 1; whereas l (u) is 0;
then, the state weight omega of the mth set of simulation data source on the d th day of the mth month is calculated by adopting the following formula1,s(d)：
Figure BDA0002914627810000075
In the formula:
Figure BDA0002914627810000076
for the indication function, K is the number of the fused precipitation data sources; when in use
Figure BDA0002914627810000077
Figure BDA0002914627810000078
On the contrary, the method can be used for carrying out the following steps,
Figure BDA0002914627810000079
and 4, considering the simulation capability of each data source on the precipitation magnitude, and calculating the dynamic magnitude weight based on the cuckoo algorithm. This step further comprises the sub-steps of:
step 4.1, optimizing memory length based on a dynamic Bayesian Mode Average (BMA) method;
in order to fuse the precipitation magnitude of the day d, the precipitation magnitude needs to be calculated by adopting each simulation data series actually measured and corrected by the station of N days before the day, and N is recorded as the memory length. In order to obtain reasonable rainfall data memory length, the rainfall data memory length is dispersed in a [31,100] interval, magnitude weights of different time sequences of each station are respectively calculated by adopting a BMA method, and then the Root Mean Square Error (RMSE) of a fusion series and an actual measurement series under each scheme is calculated:
Figure BDA00029146278100000710
in the formula: l is the total length of the actual measurement series; n represents the nth discrete series, namely a scheme with the memory length of 30+ n; sd,nAnd GdThe fusion precipitation and the actual measurement precipitation on the day d are respectively.
The BMA method firstly carries out normal conversion on meteorological site observation series and each simulation series through a Box-Cox function, and then carries out weighted average on multiple mode estimation results based on a normal linear distribution hypothesis. And for the nth set of discrete schemes, calculating the average value of the RMSE of all the stations, preferably obtaining the scheme with the minimum RMSE, and further optimizing to obtain the memory length N.
As shown in fig. 2, a schematic diagram of a posterior probability density function of a satellite precipitation data source is provided.
Step 4.2, calculating the dynamic magnitude weight of each station based on a cuckoo algorithm;
for each ground site, the minimum sum of squared deviations (SSE) is used as a target function, and the dynamic weight of each data source is calculated by adopting a cuckoo algorithm:
Figure BDA0002914627810000081
in the formula: w is as(d) Dynamic magnitude weighting of the mth month day of the simulation data source for the mth set,
Figure BDA0002914627810000082
the s set of simulation data sources representing the station j represent the precipitation magnitude corrected by the equal rate correction method in the step 2 on the day of the mth month;
Figure BDA0002914627810000083
the measured precipitation magnitude of the s-th set of simulation data sources representing the station j on day of the mth month; and N is recorded as the memory length, and K is the fused precipitation data quantity.
And 5, adopting a common Krigin interpolation method to dynamically map the correction coefficients, the state weights and the magnitude weights of all the stations to the same spatial resolution, and obtaining a long series of grid quantitative precipitation products through correction and data fusion. Step 5 further comprises the following substeps:
step 5.1, interpolating the correction coefficients and the weights of all the sites to a grid scale by adopting a common Kriging interpolation method;
interpolating the precipitation threshold and the correction coefficient of each precipitation data source in the step 2, the state weight in the step 3 and the dynamic magnitude weight in the step 4 into a grid scale; for each grid, both the state weight and the dynamic magnitude weight are normalized to give a sum of weights of 1 for each data source, wherein a spatial resolution of 0.25 ° is chosen for this embodiment.
Step 5.2, for each grid, based on the interpolated precipitation threshold value and the interpolated correction coefficient, calculating to obtain a plurality of corrected sets of simulated precipitation data sources by adopting an LOCI method and an equal rate correction method; since 5 sets of precipitation data sources are collected in step 1 of this embodiment, 5 sets of corrected precipitation data sources are obtained in step 5.2, and it is worth pointing out that for different engineering requirements, different numbers of precipitation data sources can be selected;
step 5.3, for each i grids, based on the state weight obtained by interpolation, adopting
Figure BDA0002914627810000084
Calculating the precipitation magnitude I of the ith set of simulated precipitation data source of the ith grid on the d dayi,dWherein, in the step (A),
Figure BDA0002914627810000085
simulating the precipitation magnitude of the precipitation data source on the d day for the s set of the ith grid corrected in the step 5.2, wherein K is the number of the fused precipitation data;
step 5.4, for each i grids, based on magnitude weight obtained by interpolation, adopting
Figure BDA0002914627810000091
Calculating the precipitation magnitude of the ith set of simulated precipitation data source of the ith grid on the d day
Figure BDA0002914627810000092
In this embodiment, since 5 sets of precipitation data sources are collected in step 1, K is 5, so the precipitation magnitude
Figure BDA0002914627810000093
Comprises the following steps:
Figure BDA0002914627810000094
step 5.5, adopting the precipitation magnitude I of step 5.3i,dJudging whether precipitation event occurs, if Ii,dIf < 1, precipitation eventWhen the precipitation does not occur, the fused precipitation value is determined as Ii,d(ii) a Otherwise, if the precipitation event occurs, the fusion precipitation value is calculated in the step 5.4
Figure BDA0002914627810000095
While the invention has been described with reference to a preferred embodiment, it will be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the invention.

Claims (5)

1. A multi-source precipitation product fusion method based on a dynamic heuristic algorithm is characterized by comprising the following steps:
step 1, collecting limited site observation data of a scarce data area, performing satellite inversion and re-analyzing a precipitation data set;
step 2, correcting the data collected in the step 1 by adopting a local magnitude scaling method and an equal rate correction method, and obtaining correction coefficients of all stations;
step 3, describing the capturing capacity of each data source to precipitation events through the assigning function, and calculating the state weight with space-time dynamic attributes;
step 4, considering the simulation capability of each data source on the precipitation magnitude, and calculating the dynamic magnitude weight based on the cuckoo algorithm;
step 5, adopting a common Krigin interpolation method, dynamically mapping correction coefficients, state weights and magnitude weights of all stations to the same spatial resolution, and obtaining a long series of grid quantitative precipitation products through correction and data fusion;
the specific method of the step 3 comprises the following steps:
the space-time heterogeneity of the precipitation events is fully considered, and a scoring function is adopted
Figure FDA0003545999440000016
Characterizing the detection capability of each corrected analog data source to precipitation events：
Figure FDA0003545999440000011
In the formula:
Figure FDA0003545999440000012
the rainfall magnitude corrected by using the LOCI method on the d th day of the m month by the s set of simulation data sources representing the station j is further corrected by using an equal rate correction method;
Figure FDA0003545999440000017
the measured precipitation magnitude of the ith set of simulation data source of the characterization station j in the mth month and the d day; roCharacterizing a precipitation threshold; l (u) is an indicator function, and when u is more than or equal to 0, L (u) is 1; whereas l (u) is 0;
then, the state weight omega of the mth set of simulation data source on the d th day of the mth month is calculated by adopting the following formula1,s(d)：
Figure FDA0003545999440000013
In the formula:
Figure FDA0003545999440000014
for the indication function, K is the number of the fused precipitation data sources; when in use
Figure FDA0003545999440000015
Figure FDA0003545999440000021
On the contrary, the method can be used for carrying out the following steps,
Figure FDA0003545999440000022
step 4 further comprises the following substeps:
4.1, optimizing to obtain a memory length based on a dynamic Bayesian mode averaging method;
step 4.2, calculating dynamic magnitude weight of each station based on a cuckoo algorithm; the method comprises the following steps:
for each ground station, the minimum sum of squared deviations is used as a target function, and the cuckoo algorithm is adopted to calculate the dynamic magnitude weight of each data source:
Figure FDA0003545999440000023
Figure FDA0003545999440000024
in the formula: w is as(d) Simulating the magnitude weight of the data source dynamics at day d of the mth month for the mth set,
Figure FDA0003545999440000025
the s set of simulation data sources representing the station j represent the precipitation magnitude corrected by the equal rate correction method in the step 2 on the day of the mth month;
Figure FDA0003545999440000026
the measured precipitation magnitude of the s-th set of simulation data sources representing the station j on day of the mth month; and N is recorded as the memory length, and K is the fused precipitation data quantity.
2. The dynamic heuristic algorithm-based fusion method of multi-source precipitation products of claim 1, wherein step 2 further comprises the following sub-steps:
step 2.1, correcting the precipitation occurrence frequency of each data source month by adopting a local magnitude scaling method according to the data of each ground station;
and 2.2, calculating correction factors of different months of each station based on an equal rate correction method, and obtaining the precipitation magnitude corrected by each data source.
3. The dynamic heuristic algorithm-based fusion method of multi-source precipitation products according to claim 2, wherein the specific operation method of step 2.1 is as follows:
firstly, calculating precipitation occurrence frequency of each station based on measured data, and substituting the frequency into a frequency distribution function of each simulated precipitation magnitude, so as to obtain a simulated threshold of each data source; finally, by analyzing the difference between the simulation threshold and the judgment threshold and adopting a magnitude scaling method, the precipitation occurrence probabilities of the simulation series and the actual measurement series in different months are consistent, and the specific method comprises the following steps: when the simulation threshold is larger than the judgment threshold, correcting precipitation magnitude exceeding the judgment threshold but smaller than the simulation threshold as the judgment threshold in magnitude scaling; when the simulation threshold is smaller than the judgment threshold, the precipitation magnitude lower than the judgment threshold but higher than the simulation threshold is corrected as the judgment threshold in magnitude scaling.
4. The dynamic heuristic algorithm-based fusion method of multi-source precipitation products according to claim 2, wherein the specific operation method of step 2.1 is as follows:
for the jth ground station, note
Figure FDA0003545999440000031
And calculating the correction factor of the mth month for the simulation data source of the mth set by adopting the following formula:
Figure FDA0003545999440000032
in the formula: d represents the total observed days of the mth month of the station j, and D represents the time sequence of the daily precipitation series;
Figure FDA0003545999440000033
the measured precipitation magnitude of the site j on day d of the mth month;
Figure FDA0003545999440000034
simulation of the s-th set for station jThe data source adopts the precipitation magnitude corrected by the local magnitude scaling method on the d day of the mth month;
by adopting the correction factor, the corrected magnitude of the s-th set of analog data source can be obtained as follows:
Figure FDA0003545999440000035
in the formula:
Figure FDA0003545999440000036
and (4) representing the precipitation magnitude corrected by the local magnitude scaling method on the d th day of the m month by the s set of simulation data source of the station j, and further correcting the precipitation magnitude by using an equal rate correction method.
5. The dynamic heuristic algorithm-based fusion method of multi-source precipitation products according to claim 1, wherein the specific method in step 5 is as follows:
step 5.1, adopting a common Kriging interpolation method to interpolate the correction coefficients of all the stations in the step 2, the state weight in the step 3 and the magnitude weight in the step 4 into a grid scale;
step 5.2, for each grid, based on the interpolated precipitation threshold value and the correction coefficient, adopting a local magnitude scaling method and an equal rate correction method to obtain a plurality of corrected sets of simulated precipitation data sources;
step 5.3, for each i grids, based on the state weight obtained by interpolation, adopting
Figure FDA0003545999440000037
Calculating the precipitation magnitude I of the ith set of simulated precipitation data source of the ith grid on the d dayi,dWherein
Figure FDA0003545999440000038
Simulating the precipitation magnitude of the precipitation data source on the d day for the s set of the ith grid corrected in the step 5.2, wherein K is the number of the fused precipitation data;
step 5.4, for each i grids, based on magnitude weight obtained by interpolation, adopting
Figure FDA0003545999440000039
Calculating the precipitation magnitude of the ith set of simulated precipitation data source of the ith grid on the d day
Figure FDA0003545999440000041
Step 5.5, adopting the precipitation magnitude I of step 5.3i,dJudging whether precipitation event occurs, if Ii,dIf the rainfall is less than 1, the precipitation event does not occur, and the fused precipitation value is determined as Ii,d(ii) a Otherwise, if the precipitation event occurs, the fusion precipitation value is calculated in the step 5.4
Figure FDA0003545999440000042
CN202110096858.2A 2021-01-25 2021-01-25 Multi-source precipitation product fusion method based on dynamic heuristic algorithm Active CN112766580B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110096858.2A CN112766580B (en) 2021-01-25 2021-01-25 Multi-source precipitation product fusion method based on dynamic heuristic algorithm

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110096858.2A CN112766580B (en) 2021-01-25 2021-01-25 Multi-source precipitation product fusion method based on dynamic heuristic algorithm

Publications (2)

Publication Number Publication Date
CN112766580A CN112766580A (en) 2021-05-07
CN112766580B true CN112766580B (en) 2022-04-29

Family

ID=75707088

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110096858.2A Active CN112766580B (en) 2021-01-25 2021-01-25 Multi-source precipitation product fusion method based on dynamic heuristic algorithm

Country Status (1)

Country Link
CN (1) CN112766580B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8594936B1 (en) * 2008-12-31 2013-11-26 The Weather Channel, Llc Providing current estimates of precipitation accumulations
CN110118982A (en) * 2019-04-12 2019-08-13 大连理工大学 A kind of satellite precipitation data bearing calibration based on space optimization interpolation
CN111078678A (en) * 2019-12-18 2020-04-28 中国气象局乌鲁木齐沙漠气象研究所 Satellite precipitation data correction method based on multi-source information fusion and scale reduction
CN111898660A (en) * 2020-07-17 2020-11-06 武汉大学 Hydrological simulation method for averagely fusing multi-source data based on Bayesian mode

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190018918A1 (en) * 2017-07-11 2019-01-17 The Trustees Of Princeton University System and method for performing accurate hydrologic determination using disparate weather data sources

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8594936B1 (en) * 2008-12-31 2013-11-26 The Weather Channel, Llc Providing current estimates of precipitation accumulations
CN110118982A (en) * 2019-04-12 2019-08-13 大连理工大学 A kind of satellite precipitation data bearing calibration based on space optimization interpolation
CN111078678A (en) * 2019-12-18 2020-04-28 中国气象局乌鲁木齐沙漠气象研究所 Satellite precipitation data correction method based on multi-source information fusion and scale reduction
CN111898660A (en) * 2020-07-17 2020-11-06 武汉大学 Hydrological simulation method for averagely fusing multi-source data based on Bayesian mode

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
Big Data Analysis Based on POT Method for Design Flood Prediction;Dan Li;《IEEE》;20160731;全文 *
Intercomparison of multiple statistical methods in post-processing ensemble precipitation and temperature forecasts;Xiangquan Li;《Meteorological Applications》;20201231;全文 *
Probabilistic forecasting based on ensemble forecasts and EMOS method for TGR inflow;Yixuan ZHONG;《RESEARCH ARTICLE》;20201231;全文 *
基于地理加权回归克里金的降水数据融合及其在水文预报中的应用;李运龙;《长江流域资源与环境》;20170930;全文 *
基于贝叶斯模式平均方法融合多源数据的水文模拟研究;尹家波;《水利学报》;20201130;全文 *

Also Published As

Publication number Publication date
CN112766580A (en) 2021-05-07

Similar Documents

Publication Publication Date Title
CN107316095B (en) Regional weather drought level prediction method coupled with multi-source data
CN111898660B (en) Hydrological simulation method for average fusion of multi-source data based on Bayesian mode
Stathopoulos et al. Wind power prediction based on numerical and statistical models
CN111027175B (en) Method for evaluating social and economic influences of flood based on coupling model integrated simulation
CN114019579B (en) High space-time resolution near-surface air temperature reconstruction method, system and equipment
Donat et al. High-resolution refinement of a storm loss model and estimation of return periods of loss-intensive storms over Germany
CN107403073B (en) Integrated flood forecasting method for improving and forecasting rainfall based on data assimilation
CN115507822B (en) Flood risk prediction method driven by hydrologic cycle variation
CN113158542B (en) Multivariable design flood estimation method suitable for data-lacking area
KR101383617B1 (en) Method and apparatus for predicting daily solar radiation level
CN107609713B (en) Land-air coupling real-time forecasting method through rainfall and runoff double-factor correction
CN110598290A (en) Method and system for predicting future hydropower generation capacity of basin considering climate change
Kim et al. Statistical downscaling for daily precipitation in Korea using combined PRISM, RCM, and quantile mapping: Part 1, methodology and evaluation in historical simulation
CN112700068B (en) Reservoir dispatching rule optimization method based on machine learning fusion of multi-source remote sensing data
CN113553782B (en) Downscaling method for forecasting wind speed
CN114254802B (en) Prediction method for vegetation coverage space-time change under climate change drive
KR20140103589A (en) Flood estimation method using MAPLE forecasted precipitation data and apparatus thereof
CN110909447B (en) High-precision short-term prediction method for ionization layer region
Tousi et al. Climate changes impact on stormwater infrastructure design in Tucson Arizona
CN108808671A (en) A kind of short-term wind speed DATA PROCESSING IN ENSEMBLE PREDICTION SYSTEM method of wind power plant
CN117172037B (en) Distributed hydrologic forecasting method, device, computer equipment and medium
CN112766580B (en) Multi-source precipitation product fusion method based on dynamic heuristic algorithm
CN113392365A (en) High-resolution meteorological grid data generation method and system
CN117493476A (en) Runoff backtracking simulation method and system integrating physical mechanism and artificial intelligence
AU2021105536A4 (en) A High Spatial-Temporal Resolution Method for Near-Surface Air Temperature Reconstruction

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant